structural and functional characterisation of idia/futa (tery ......structure and function of...

Structure and Function of Trichodesmium Tery_3377

Structural and functional characterization of IdiA/FutA (Tery_3377), an iron binding protein from theocean diazotroph Trichodesmium erythraeum

Despo Polyviou‡,1, Moritz M. Machelett§, ‡,1, Andrew Hitchcock‡,1,2, Alison J. Baylay‡, FraserMacMillan¶, C. Mark Moore‡, Thomas S. Bibby‡ and Ivo Tews§,3

From the ‡Ocean and Earth Sciences, National Oceanography Centre Southampton, University ofSouthampton, Waterfront Campus, European Way, Southampton SO14 3ZH, UK, the §Department of

Biological Sciences, Faculty of Natural and Environmental Sciences Life Sciences, Building 85, Universityof Southampton, Highfield Campus, Southampton SO17 1BJ, UK, and the ¶School of Chemistry,

University of East Anglia, Norwich Research Park, Norwich NR4 7TJ, UK 1The first authors should beregarded as Joint First Authors. 2Present address: Department of Molecular Biology and Biotechnology,

University of She�eld, Firth Court, Western Bank, She�eld S10 2TN, UK.

Running title: Structure and Function of Trichodesmium Tery_3377

To whom correspondence should be addressed: 3 Ivo Tews, Tel.: +44 (0)23-8059-4415; E-mail: [email protected].

Keywords: cyanobacteria, ABC transporter, crystal structure, microscopy, Trichodesmium, iron, IdiA, FutA,diazotroph, iron deficiency

ABSTRACTAtmospheric nitrogen fixation by photosyn-

thetic cyanobacteria (diazotrophs) strongly influ-ences oceanic primary production and in turn af-fects global biogeochemical cycles. Species of thegenus Trichodesmium are major contributors to ma-rine diazotrophy, accounting for a significant propor-tion of the fixed nitrogen in tropical and subtropicaloceans. However, Trichodesmium spp. are metabol-ically constrained by the availability of iron, an es-sential element for both the photosynthetic appara-tus and the nitrogenase enzyme. Survival strategiesin low-iron environments are typically poorly char-acterized at the molecular level, as these bacteriaare recalcitrant to genetic manipulation. Here, westudied a homolog of the iron deficiency-inducedA (IdiA)/ferric uptake transporter A (FutA) pro-tein, Tery_3377, which has been used as an in situiron-stress biomarker. IdiA/FutA has an ambigu-ous function in cyanobacteria, with its homologshypothesized to be involved in distinct processes de-pending on their cellular localization. Using sig-nal sequence fusions to GFP and heterologous ex-pression in the model cyanobacterium Synechocys-tis sp. PCC 6803 we show that Tery_3377 is tar-

geted to the periplasm by the twin-arginine translo-case and can complement the deletion of the nativeSynechocystis ferric-iron ABC transporter periplas-mic binding protein (FutA2). EPR spectroscopyrevealed that purified recombinant Tery_3377 hasspecificity for iron in the Fe3+ state, and an X-raycrystallography determined structure uncovererd afunctional iron substrate-binding domain, with Fe3+penta-coordinated by protein and bu�er ligands. Ourresults support assignment of Tery_3377 as a func-tional FutA subunit of an Fe3+ ABC-transporter, butdo not rule out that it also has dual IdiA function.

Photosynthetic organisms have a higher requirementfor iron than heterotrophs because of the dependencyof multiple components of the photosynthetic elec-tron transport chain on this essential element (1,2). This requirement is amplified in nitrogen fixingcyanobacteria (diazotrophs) as Fe is also required bythe nitrogenase enzyme (3, 4, 5). Therefore, marinediazotrophic cyanobacteria species such as the glob-ally significant Trichodesmium have developed di-verse strategies that permit survival at the fluctuatingor chronically depleted iron concentrations typically

1

http://www.jbc.org/cgi/doi/10.1074/jbc.RA118.001929The latest version is at JBC Papers in Press. Published on September 14, 2018 as Manuscript RA118.001929

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/cgi/doi/10.1074/jbc.RA118.001929http://www.jbc.org/


encountered in the surface waters of the tropical andsubtropical oceans (5, 6, 7, 8, 9). Trichodesmiumaccounts for approximately half of the total marinenitrogen fixation (10), supplying both biologicallyavailable nitrogen and fixed carbon to oligotrophicecosystems (11). Physiological experiments suggestTrichodesmium can acquire Fe from diverse sources(12, 7) and analysis of its genome reveals homologsto Fe-uptake systems, such as FeoAB for ferrous(Fe2+) iron uptake and the periplasmic ferric hydrox-amate binding protein FhuD (5). While ’omic’ stud-ies and similarity to homologous systems in otherbacteria can be used to infer function, the inabilityto genetically engineer Trichodesmium means no Feuptake mechanisms have been directly characterizedin vivo, limiting our understanding of how this im-portant species has adapted to acquire Fe from itsenvironment.

Ubiquitous in cyanobacterial genomes andpresent in Trichodesmium are homologs of a proteinvariably referred to in the literature as IdiA (iron de-ficiency induced) or FutA (ferric uptake transporter).IdiA/FutA-like proteins are synthesized under Felimitation in marine and freshwater cyanobacteria(13, 14), and as such are commonly exploited as insitu markers of Fe stress (15, 16, 17). Despite this,functional assignment of these proteins is presentlyambiguous. homologs can have an intracellular lo-calization (18) and are proposed to be associatedwith the cytoplasmic side of the thylakoid membraneto protect the acceptor side of photosystem II (PSII)against oxidative damage during Fe stress (18, 19).However, the same protein is also homologous toferric binding proteins (FBPs) of ATP-binding cas-sette (ABC) transporters, found in some pathogenicbacteria and several cyanobacteria (20, 21, 22).

In the model freshwater cyanobacteriumSynechocystis sp. PCC 6803 (hereafter Syne-chocystis), there are two characterized paraloguesof IdiA/FutA, referred to as FutA1 (Slr1295) andFutA2 (Slr0513). FutA1 has a predominantly intra-cellular localization, associating with the thylakoidmembrane, although it has also been found in thecytoplasmic membrane. In contrast, the FutA2 pro-tein is a periplasmic substrate binding protein (SBP)associated with partner FutB and FutC subunits toform an active Fe3+ uptake ABC transporter (22, 23,24, 25, 26).

Trichodesmium erythraeum IMS101 en-codes only one IdiA/FutA homolog in its genome,Tery_3377 (17). Increased expression/productionunder Fe stress has been reported at both thetranscriptional (3, 17), and protein levels (5, 13).Tery_3377 is therefore thought to play an importantrole in iron homeostasis.

In this study we investigated the role ofTery_3377 to understand whether it is performing aperiplasmic function (FutA2-like) or an intracellularfunction (FutA1-like). We produced Tery_3377 infutA1 and futA2 mutants in the genetically tractablemodel cyanobacterium Synechocystis sp. PCC 6803testing for complementation of mutant phenotypes.We also used signal peptide fusions to GFP to giveinsight into the cellular localization of Tery_3377.Further, we overproduce and purify recombinantTery_3377 and confirm ferric iron (Fe3+) coordina-tion by EPR spectroscopy and structure determina-tion, classifying Tery_3377 as a class IV Fe substratebinding domains (27). This functional and structuralinvestigation of Tery_3377 helps to characterize itsrole in the iron-stress physiology of a globally sig-nificant microbe and suggests a potential need toreassess the roles of FutA/IdiA family proteins moregenerally.

RESULTSInsights from sequence analysis

The protein sequence of the Tery_3377protein is 58% and 57% identical to Synechocys-tis FutA1 and FutA2, respectively (Supplemen-tary Information, Table S1). In Synechocystis,FutA1 and FutA2 are 52% identical to one an-other (21), yet appear to have distinct cellular func-tion. Functional assignment based on primary se-quence comparisons is therefore not possible. Inaddition, like the Synechocystis futA1 and futA2genes, the Tery_3377 gene is located at an inde-pendent locus to genes encoding putative Fut ABC-transporter permease (FutB/Tery_3223) and ATPasesubunits (FutC/Tery_3222), making functional in-ference from genomic context impossible.

Phylogenetic analysis was carried out to in-vestigate grouping of Tery_3377 with IdiA/FutA ho-mologs (Figure 1). A multiple sequence alignmentwas constructed from Tery_3377, FutA1 and FutA2,along with homologs from range of cyanobacterial

2

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


species, and this was used to generate a maximumlikelihood phylogenetic tree. The SynechocystisFutA1 and FutA2 proteins are separated into dis-tinct branches within the tree (branch support 0.97,approximate likelihood ratio test). Tery_3377 ap-pears to be more closely related to the intracel-lular FutA1 protein than to FutA2. To confirmplacement of the Tery_3377 protein within the phy-logeny, we searched for homologs in other publiclyavailable Trichodesmium genomes / metagenomes.No Tery_3377 homolog could be detected in thedraft genome of T. theibautii H94 (Bioproject PR-JNA278659); however this genome is known to beincomplete (28). A homolog was found in a metage-nomic dataset from the Bermuda Atlantic Time Se-ries (Bioproject PRJNA336846). This protein (en-coded between nucleotides 342 and 1392 on contig00877) clustered with Tery_3377 (Figure 1).The Tery_3377 signal peptide targets GFP to theperiplasmTo investigate the cellular localization of Tery_3377a synthetic gene encoding the predicted signal pep-tide (aa 1-31) and first 4 residues of the matureprotein (aa 32-35) fused to super-folder (sf)GFPwas expressed in Synechocystis under the controlof the psbAII promoter (Table S1 and Figure S1).For comparison, the predicted Synechocystis FutA1and FutA2 signal peptides were also fused to sfGFPand expressed from the psbAII promoter. The Es-cherichia coli TMAO reductase (TorA) signal pep-tide that has previously been shown to transportGFP/YFP into the periplasm of Synechocystis wasused as a positive control, and sfGFP alone, i.e.without signal sequence, acted as a marker of cyto-plasmic expression (29 , 30).

The strains were visualized by confocal mi-croscopy. Excitation of WT cells at 488 nm does notresult in fluorescence emission in the GFP channel at510-530 nm, but prominent auto-fluorescence fromthylakoid membrane localized photosynthetic pig-ments was observed when emission was monitoredat 600-700 nm (Figure 2A), as expected. Fusion ofsfGFP to the TorA (Figure 2B) or FutA2 (Figure 2D)signal peptide results in bright fluorescent halos thatsurround the auto-fluorescence signal in merged im-ages, indicating translocation of GFP across the cy-toplasmic membrane to the periplasm. It is perhapssurprising that the FutA1 signal peptide also appears

to localize some GFP to the periplasm or cytoplas-mic membranes, albeit to a lesser extent than withthe the TorA or FutA2 fusions (Figure 2C). Previousstudies have suggested that FutA1 is mainly associ-ated with thylakoid membranes, although it has beendetected in the plasma membrane (26).

Fusion of the Tery_3377 signal peptide toGFP results in distinct halos (Figure 2E), althoughthis is not as clear as for the TorA and FutA2 signalpeptides (Figures 2B and 2D), and also showed sim-ilarity to the FutA1 fusion (Figure 2C). We changedthe twin arginine residues in the Tery_3377 sig-nal peptide to lysines, which prevented export ofGFP and showed that the wild-type Tery_3377 sig-nal peptide is recognized and exported by the twin-arginine translocase (TAT) system (Figures S2). Im-munoblots with an anti-GFP antibody identified aprotein approximately 3 kDa larger than sfGFP in thestain with the mutated signal peptide, consistent withthe presence of the uncleaved signal peptide (FigureS3). In the strain producing just sfGFP without asignal peptide clear fluorescence signal mapping tothe cytoplasm was observed (Figure 2F), as expectedand as reported previously for YFP (30).

These results indicate that the Tery_3377signal peptide can translocate GFP to the periplasmusing the TAT system. Both of the native Syne-chocystis FutA signal peptides also result in somedegree of cytoplasmic membrane or periplasmic lo-calization of GFP. Based on previous work FutA1 isconsidered to be predominantly intracellular, hencethis result is unexpected. It is likely that the matureprotein sequences also play a role in cellular local-ization, but attempts to investigate this in strainsin which GFP was fused to the full proteins ratherthan just the signal peptides were unsuccessful asthe strains were either genetically unstable or GFPwas cleaved from the protein (data not shown).Tery_3377 is involved in Fe3+ transport across thecell membraneThe signal peptide fusions indicated that Tery_3377is at least partly localized in the periplasm, pre-sumably functioning as the periplamic binding pro-tein component of a ferric uptake ABC-transporter.To gain further insight into the functional role ofTery_3377, the gene was codon optimized for ex-pression in Synechocystis and integrated at the ps-bAII locus in �futA1 and �futA2 Synechocystis mu-

3

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


tants. Production of Tery_3377 was confirmed byimmunoblotting with the anti-IdiA antibody (raisedagainst Synechococcus elongatus PCC 6302 IdiA),but was not detected using an antibody against Syne-chocystis FutA2 (Figure S4).

The high a�nity iron ligand ferrozine (FZ)scavenges Fe2+ and cannot penetrate the cell mem-brane and thus is hypothesized to cause a Fe stressphenotype through reduction of ferrous iron uptakethrough the FeoAB system (31, 32, 33). This ef-fect would be expected to be more severe in strainswhere ferric iron uptake is also reduced, such as the�futA2 strain. Experiments were performed in bio-logical triplicate and di�erences between FZ treatedand non-treated cells for each strain were determinedfor growth and photosynthetic e�ciency (Figure 3).A reduction in photosynthetic e�ciency (Fv/Fm) isexpected under iron stress because of the essentialrole of iron in the photosynthetic electron transportchain (4, 5).

In the presence of FZ, growth rates weresignificantly reduced by 13% in the WT (P < 0.01)and 15% in the �futA1 strain (P < 0.05, Figure 3A)while the �futA2 mutant lacking the Fe3+ trans-porter showed the greatest response, as expected(32% reduction, P < 0.01). The reduction in growthrate by FZ in both mutants was not as apparentwhen Tery_3377 was introduced to complement theknockout strains �futA1 and �futA2, suggesting thatTery_3377 enables the cells to acquire ferric ironwhen ferrous iron is bound by FZ. Growth in pres-ence of FZ also caused a reduction in photosynthetice�ciency (Fv/Fm) that was largest for�futA2 (34%).The negative e�ect on the �futA2 phenotype wasagain alleviated in the tery_3377 complementationstrain (Figure 3B). In contrast, the reduction in pho-tosynthetic e�ciency in the �futA1 knockout is lesssevere than WT.Tery_3377 binds ferric (Fe3+) ironRecombinant Tery_3377 with a C-terminal His-tagwas over-produced in E. coli and purified in orderto investigate its a�nity and binding of Fe species.After incubation with Ammonium Fe(II) sulfate un-der aerobic conditions, the Tery_3377-Fe complexwas isolated by size exclusion chromatography to re-move unbound Fe. The purified protein-Fe complexwas analyzed by Electron Paramagnetic Resonance(EPR) spectroscopy to determine the oxidation state

of the bound Fe ligand. EPR spectra revealed anoxidized Fe3+ state present in the Tery_3377 sample(Figure 4). In addition to weaker signals between800 – 1400 Gauss, a sharp and intense signal atabout 1500 Gauss (equal to a g-value of 4.3) wasclearly observed. This g-value (4.3) is indicative ofa high-spin Fe3+ ion bound to Tery_3377. The typ-ical fingerprint weaker signals are related to higherorder spin transitions of this ferric iron.Crystal structures of Tery_3377 reveal penta-coordination of Fe3+To study Fe3+ binding by Tery_3377, the structureof the iron-bound protein was determined using X-ray crystallography (Table 1). The overall fold ofTery_3377 assumes a classic small molecule ABCtransporter substrate-binding domain fold, with twodomains hinged over a substrate-binding cleft. Thehinge-region is formed by two short beta-strands andconnects the protein domains via a double crossoverbetween the N- (residues 34-133 and 266-349) andC-terminus (134-265). This type of structure placesTery_3377 amongst Class IV Fe-binding proteins(in keeping with 34), and allows for structural clas-sification within Cluster D of the substrate-bindingdomains (35, 36).

Iron substrate binding domains of di�erentclasses di�er in Fe ligand coordination, determinedby conserved interacting amino acid side chains ofthe substrate binding domain, often involving tyro-sine residues. In Tery_3377 three tyrosine residuesTyr144, Tyr200 and Tyr201 from the C-terminaldomain contribute to Fe binding. In the struc-ture of Tery_3377-Ala (PDB: 6G7N), determinedat 1.1 Å resolution, a bu�er molecule from the crys-tallization mixture occupies two coordination sites,giving rise to a penta-coordinated Fe assigned asFe3+ (Figure 5A). The geometry observed is trig-onal bi-pyramidal, with the bu�er molecule ala-nine carboxyl-group, Tyr174 and Try231 formingthe trigonal plane with angles of 108°, 108° and144°. Tyr230 and the amino-group from the bu�eralanine coordinate the axial Fe positions.

Amino acid additives in the crystallizationbu�ers lead to the most ordered and best di�ractingcrystals. However, Tery_3377 crystallized undera variety of conditions, including ones that lackedamino acid additives. We determined the struc-ture of the protein crystallized with a monosaccha-

4

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


ride additive mixture and found it was devoid of anFe-coordinating ligand in the binding site, with theiron instead coordinated by ordered water molecules(Figure 5B). Comparison of the Tery_3377-Waterstructure (PDB: 6G7P) to the Tery_3377-Ala com-plex shows that water molecules W1 and W2 takethe positions of the alanine amino and carboxylategroups of the alanine ligand, respectively (Figure5). Interestingly, in this complex an additional wa-ter molecule ligand (W3) is inserted between thecoordinating oxygen atoms of W2 and Tyr231. Incomparison to Tery_3377-Ala, where the observedangle between Tyr23-1O and Alanine-carboxyl oxy-gen was measured as 144°, this angle is split into twoangles of roughly 73° measured between Tyr231-O,the inserted water molecule W3, and W2. Like inTery_3377-Ala, Tyr230 and water W1 occupy theaxial coordination sites in Tery_3377-Water.

To try and crystalline ligand-free Tery_3377in order to examine structural changes upon ironbinding, we citrate-treated Tery_3377 protein prepa-rations prior to crystallization; citrate is knownto chelate Fe and was added to remove all Fefrom the protein preparation. The crystal struc-ture (Tery_3377-Cit, PDB: 6G7Q) traps the citratemolecule with bound Fe in the substrate bindingcleft, where the Fe is removed approximately 7 Åfrom the binding site (Figure 6). The Fe binding site(pink sphere in Figure 6) was free of metal, and thesignificance of this finding in the context of possiblesiderophore acquisition is discussed below .

DISCUSSIONWe have investigated the function and structure ofTrichodesmium Tery_3377, an Fe stress induced pro-tein (iron deficiency induced A, IdiA) commonlyused as an in situ marker of Fe stress (15, 16, 17).Phylogenetic analysis (Figure 1) groups Tery_3377with the Synechocystis FutA1 protein, which hasa suggested intracellular PSII protective function,rather than the FutA2 protein, a periplasmic proteinwith a role in Fe-transport. However, owing to thehigh level of sequence identity of Tery_3377 to bothFutA1 and FutA2 proteins, and FutA1/2 proteinsto one another, the function of these similar iron-binding proteins may be determined by their cel-lular localization. By fusion to sfGFP, Tery_3377is shown to have a functional signal sequence for

transport to the periplasm (Figure 2) by the TATsystem (Figures S1 and S2). Consistent with this,Tery_3377 complements the ferrozine (FZ) inducedFe stress phenotype of the Synechocystis �futA2deletion strain (Figure 3). These data support func-tional assignment of _3377 as a subunit of a Fe-uptake ABC transporter.

Expression of Tery_3377 also complementsthe more subtle phenotype of the �futA1 Syne-chocystis mutant in the presence of FZ. This mayindicate the possibility for an additional intracellu-lar function in PSII protection, or alternatively re-sult from increased ferric iron uptake by Tery_3377.Since localization data using FutA1 signal peptideGFP fusions resulted in some periplasmic localiza-tion (Figure 2)), FutA1 may have a role in iron uptakein addition to its reported intracellular function. In-deed, despite their apparently distinct cellular loca-tions and functions, a futA1/futA2 double mutant inSynechocystis has more severe iron-deficient growthand ferric iron uptake phenotypes than either sin-gle mutant, suggesting some degree of redundancy(20). In agreement with this, the phenotype of thedouble mutant is similar to that of mutants of thesingle copy Fut ABC-transporter permease (futB)and ATPase subunit (futC) genes (20). Hence theIdiA/FutA family of proteins may commonly displaya dual function that requires a broader reassessmentacross cyanobacteria.

EPR spectroscopy and structural analysisconfirms Tery_3377 binds Fe in the ferric oxidationstate, Fe3+ (Figure 4). Tery_3377 crystallizes undervarious conditions, often requiring small-moleculeadditives. The best di�racting crystals were obtainedwith an additive mixture of amino acids (Figure 5A).An amino acid ligand, alanine, is identified in the Febinding cleft; the amino acid is ordered and coor-dinates the Fe substrate. In contrast, other crys-tallization conditions, which did not show ligandcoordination by organic molecules, gave less well-ordered crystals that di�racted to lower resolutionand generally refined to higher R-values, indicatinga lower quality of the resulting models (Table 1).We also determined the structure where organic lig-ands are absent from the Fe binding site, with threewater molecules coordinating the Fe, in place of thetwo coordinating atoms seen with the alanine ligandstructure (Figure 5B).

5

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


To determine the most likely coordina-tion state of Fe in Tery_3377, we superposed theTery_3377-Ala structure with a classic closed con-formation FutA2 structure (Synechocystis PCC6803FutA2, PDB: 2VP1) (Figure 7A). The coordinatingoxygen of the carboxylic acid moiety of the ala-nine ligand ideally superposes with a tyrosine sidechain oxygen, while the amino nitrogen of the ala-nine ligand superimposes with a nitrogen from aninteracting histidine ligand (Figure 7B,C). The twocoordinating amino-acids in FutA2 are contributedby a loop that is brought into the vicinity of theFe binding site by domain closure around the hingethat brings the two lobes of the substrate bindingdomain closer together (37, 38) (Figure 7A). Inter-estingly, this motif is conserved in Tery_3377, andsuch a domain closure, while not observed underany of our crystallization conditions, can be pro-posed. We conclude that the alanine ligand mimicsthis Fe binding mode leading to a penta-coordinatedstate. This coordination state is corroborated by thelow-temperature solution EPR spectroscopy, whichrevealed a strong signal at g=4.3 and less intensehigher-order transitions at higher g-values, support-ing generic Class IV Fe coordination by Tery_3377in solution.

The observed binding of organic ligandstogether with Fe may just be a fortuitous obser-vation, but it might also have biological signifi-cance. It could be argued that the high concentra-tions of alanine present in crystallization mixturesout-competes interactions with the N-terminal Fe-coordination residues His43 and Tyr44, thus leadingto preferred crystallization of an open conformation.However, specific bu�er molecules such as carbon-ate have previously been shown to coordinate Fe insubstrate binding domains (39). Indeed, the bindingsite is large enough to fit a variety of amino acids insimilar fashion to the shown alanine in Tery_3377-Ala. Hence Fe binding might concur with bindingof di�erent, larger molecules in the binding site.Nitrogen containing molecules might enrich in theperiplasm, creating a concentration gradient over theouter membrane. Additional support for the plastic-ity in the binding of Tery_3377 comes from the ob-servation of bound citrate in the binding cleft, clearlyindicating the potential for binding of organic com-pounds (Figure 6). It is interesting to speculate that

binding of organic ligands together with iron mayrepresent a mechanism of siderophore recognition.SummaryThis study provides structural and mechanistic in-sight into a key protein important in iron acquisitionand the iron stress response in the globally signif-icant marine diazotroph Trichodesmium. These or-ganisms are central to N2 fixation, a highly signifi-cant process in the biogeochemical cycle, and whosebiogeography is determined by the availability ofiron. Using heterologous expression in the modelcyanobacteria Synechocystis PCC6803 we provideevidence that Tery_3377, a commonly used iron-limitation biomarker, is localized to the periplasmwith a suggested FutA2 function as an iron bindingprotein of an ABC transporter. However, Tery_3377also complements deletion of FutA1, which couldindicate it also functions in protection of PSII, sug-gesting a potential dual function of Tery_3377 andmore generally the IdlA/Fut family of proteins incyanobacteria. Structural and spectroscopic charac-terization of recombinantly produced Tery3̀377 con-firm ferric iron binding under aerobic conditions.The structural data further show a remarkable plas-ticity of the substrate binding cleft, allowing organicmolecules and potentially siderophores to bind inaddition to Fe3+.

EXPERIMENTAL PROCEDURESProtein phylogeny analysisMaximum likelihood phylogenetic analysis of pro-tein sequences of Tery_3377 homologs from Syne-chocystis and selected bacteria was performed toidentify possible functional clusters. homologs ofthe Trichodesmium Tery_3377 and two Synechocys-tis proteins FutA1 and FutA2 were identified bysearching the NR protein database with BlastP (40).The top ten closest hits were retained for each querysequence, and clustered at 90% identity using CD-HIT (41) to reduce redundancy. For each cluster,one representative was retained for inclusion in theanalysis. homologs from Prochlorococcus mari-nus MIT 9301, Synechococcus elongatus PCC 6301and PCC 7942 and Anabaena sp. PCC 7120 werealso included in the analysis as representatives ofimportant primary producers, and sequences fromHaemophilus influenza and Neisseria meningitideswere chosen to represent pathogens.

6

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


The online platform Phylogeny.fr (42, 43)was utilized in "a la carte" mode to create an anal-ysis pipeline. Sequences were firstly aligned withT-co�ee (v11.00.8cbe486, 44) and the alignmentwas curated with G-blocks (v0.91b, 45). Subse-quently, a maximum likelihood phylogenetic treewas constructed using PhyML v3.1/3.0, (46) andan approximate likelihood ratio test was carried outto compute branch supports (aLRT, 47). Finally, agraphical representation was produced with FigTreev1.4.3 (Institute of Evolutionary Biology, Universityof Edinburgh).Growth conditionsSynechocystis strains (Table S2) were routinelygrown in BG-11 media (48) supplementedwith 5 m� glucose and bu�ered to pH 8.2with 10 m� N-17 [tris(hydroxymethyl)methyl]-2-aminoethanesulfonic acid (TES)-KOH. Cultureswere maintained at 30 �C, shaking at 150 rpm underconstant illumination of ca. 50 µ� photons m�2 s�1.For growth on plates, 1.5% (w/v) agar and 0.3%(w/v) sodium thiosulphate were added. Antibioticswere included as appropriate (Table S2). Ferrozine(FZ) growth experiments were carried out in Fe(0.6 m� FeCl3) YBG-11 media (recipe describedby 49) at a FZ final concentration of 200 µ� (FZ+)or no added FZ (FZ-). Growth was assessed underphoto-autotrophic conditions (no glucose), in tripli-cate cultures illuminated by 30 µ� photons m�2 s�1of red light to minimize abiotic photoreduction ofFe in the presence of FZ (50).

Escherichia coli (Table S2) was grown shak-ing (250 rpm) at 37 �C in LB (Luria-Bertani) brothwith 100 µg ml�1 ampicillin. Plates were set with1.5% (w/v) agar.Photosynthetic physiologyPhotosynthetic e�ciency of Synechocystis strainsduring FZ growth experiments was measured asFv/Fm, an estimate of the apparent PSII photochem-ical quantum e�ciency (51), using a FASTtrackaTMMkII Fast Repetition Rate fluorometer (FRRf) inte-grated with a FastActTM Laboratory system (ChelseaTechnologies Group Ltd., Surrey, United Kingdom).Samples were dark adapted for 20 min prior to mea-surements. Presented results were calculated fromthe average of three technical replicates for each ofthree biological replicates.Generation of Synechocystis strains

In this study, (1) pET21a was used for protein pro-duction in E. coli; (2) pACYC184 was used as atemplate to amplify the chloramphenicol resistancecassette; (3) pZEO was used as a template to amplifythe zeocin resistance cassette; (4) pPD-FLAG wasused as delivery vector for the introduction of genesinto the Synechocystis genome in place of psbAII;and (5) pED151 was used for isolation of sfGFP.

For the generation of signal peptide fusionsto sfGFP, DNA sequence corresponding to the pre-dicted (in the case of futA1, futA2, Tery_3377) orknown (Escherichia coli torA; 52) signal sequenceand first 4 amino acid residues of the mature protein(as identified by SignalP v4.1/v3.0 sensitivity andTatP v1.0 servers 53, 54) (Table S1, Figure S1) wasamplified from the respective organisms genomic(g)DNA (for details of all primers used in this studysee Table S3). Approximately 400 bp immediatelypreceding the psbAII gene was joined to each sig-nal sequence by overlap-extension (OLE)-PCR suchthat the ATG start codon of the signal sequence re-placed the ATG start codon of psbAII. Using thesame technique the chloramphenicol acetyl trans-ferase (cat) gene from pACYC184 (NEB, UK) wasinserted between the coding sequence for sfGFP (55)and 4̃00 bp of sequence immediately downstreamof psbAII. Finally, each psbAII upstream-signal se-quence fragment was joined to the sfGFP-cat-psbAIIdownstream fragment, again using OLE-PCR. ThepsbAII locus is extensively utilized as a ‘neutral site’that will not result in polar e�ects following geneinsertion, and is commonly used for expression ofrecombinant genes (56).

To generate a variant of the Tery_3377 sig-nal peptide in which the twin arginine residues werereplaced by lysines (3377ss-KK-GFP) alternativeprimers were designed to anneal during OLE-PCRand both contained the first 18 bp of the Tery_3377signal sequence with changed nucleotide sequencebetween positions 13-18 (AAA AAA in the place ofAGA CGA) such that the translated sequence codesfor two lysine residues in the place of the two argi-nine residues of the twin arginine transport (TAT)system.

Synechocystis was transformed as de-scribed by 57 with selection on chloramphenicol(8.5 µg ml�1). Genome copies were segregated byrepeated streaking on BG11 plates with increas-

7

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


ing antibiotic concentration up to 34 µg ml�1 andhomozygous transformants were verified by PCRand sequencing (Eurofins MWG Operon, Ebersberg,Germany).

For the generation of slr1295 (�futA1) andslr0513 (�futA2) Synechocystis deletion mutants,the antibiotic resistance cassette for zeocin (ZeoR)from pZEO (58) or the cat gene were used. PCRproducts consisting of the sequence upstream of thefutA1/futA2 genes and the beginning of the openreading frame (ORF), or the end of the ORF anddownstream flanking DNA, were amplified fromSynechocystis gDNA. Using (OLE)-PCR, ZeoR orcat was inserted between the upstream and down-stream PCR products to create a linear mutagenesisfragment. For slr1295 (futA1) the entire gene apartfrom the first 22 bp was replaced with the zeo cas-sette. For slr0513 (futA2) all but the first 66 bp andlast 41 bp of the gene were replaced by the cat cas-sette. The deletion constructs were introduced intoWT Synechocystis as described above, with selec-tion and segregation on zeocin (2.5-20 µg ml�1) orchloramphenicol (8.5-34 µg ml�1). Full segregationof mutant strains was confirmed by PCR.

For expression of tery_3377 in Synechocys-tis, the gene was codon optimized (Integrated DNATechnologies Inc, Leuven, Belgium) and cloned intothe NdeI/BglII sites of pPD-FLAG (56). Trans-formation of Synechocystis with the constructedplasmid integrates tery_3377 into the Synechocystisgenome such that expression is driven from the ps-bAII promoter, and confers kanamycin resistance al-lowing selection of transformed cells using increas-ing concentrations of kanamycin (10-40 µg ml�1).Segregation of transformants at the psbAII locus wasconfirmed by PCR and the sequence of the gene wasverified by sequencing, as above.Fluorescence microscopySynechocystis cells producing sfGFP fusion proteinswere visualised using a Leica SP5 LSCM confo-cal microscope (Leica Microsystems, Wetzlar, Ger-many) or a axioscope Z plus epifluorescence micro-scope (Zeiss, Oberkochen, Germany). For confocalmicroscopy, samples were excited at 488 nm andemission was measured at 510-530 nm for sfGFPfluorescence or 600-700 nm for autofluorescence, asdescribed by 59. For visualization of sfGFP by epi-fluorescence microscopy, excitation was set at 400-

490 nm (max 441 nm) and emission was measuredat 460-570 nm (max 485 nm).Electron paramagnetic resonanceOxidised Tery_3377 (550 µ�) was placed in EPRquartz tubes (Wilmad) and shock-frozen in liquidnitrogen. X-band continuous wave (cw) EPR spec-tra were recorded on a Bruker eleXsys E500 spec-trometer using a standard rectangular Bruker EPRcavity (ER4102T) equipped with an Oxford heliumflow cryostat (ESR900). The spectrometer operatesat X-band (about 9.4 GHz) frequency and uses a 1mT modulation amplitude (peak to peak) and 2 mWmicrowave power. Spectra were recorded at cryo-genic temperatures (5-6 K). A typical spectrum wasrecorded in about 15-30 mins.Protein production, crystallisation and structuredeterminationTo produce C-terminally His6-tagged Tery_3377 inE. coli, the open reading frame (minus the puta-tive sequence encoding the N-terminal signal pep-tide and stop codon) was amplified from the Tri-chodesmium genome. The resulting PCR productwas digested and cloned into the NdeI/XhoI sitesof pET21a(+) (Novagen). The sequence verifiedplasmid was introduced into E. coli BL21(DE3)and production of the recombinant protein was in-duced in cells grown to an OD600 of 0.6 by addi-tion of 0.4 m� isopropyl-�-D-thiogalactopyranoside(IPTG), followed by growth for a furher 4 hours.Cells were harvested by centrifugation (6000 x g,20 min, 4 �C), resuspended in binding bu�er (50 m�Tris, 200 m� NaCl, 5% Glycerol, 20 m� Imoda-zole), broken (sonication for 10 sec on, 30 seco�; 2 min process time) and the cell-free extractwas isolated as the supernatant following centrifu-gation (200000 x g, 45 min, 4 �C). His-taggedTery_3377 was purified by a�nity chromatography(1 ml HisTrap HP columns, GE Healthcare) and size-exclusion chromatography (HiLoad 16/60 Superdex200, GE Healthcare).

Tery_3377 crystals were obtained in spacegroup P21 with two molecules in the asymmetric unitwith slightly di�erent packing between Fe boundand Fe free forms of Tery_3377. Hanging dropvapor di�usion used conditions based on or opti-mized from the Morpheus screen (Molecular Di-mensions). The reservoir composition leading toTery_3377-Ala crystals was 0.1 � Amino acids Mix,

8

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


0.1 � Bu�er System 2, pH 7.5, 50 % PrecipitantMix 1. Tery_3377-Ala data were collected at theDiamond Light Source beamline I02 to a resolutionof 1.1 Å. Tery_3377-Water crystals were obtainedwith a reservoir composition of 0.12 � Monosac-charides Mix, 0.1 � Bu�er System 2, pH 7.5, 50 %Precipitant Mix 1. Tery_3377-Water data were col-lected at the Diamond Light Source beamline I02 toa resolution of 1.5 Å. Tery_3377-Cit crystals wereobtained with a reservoir composition of 0.1 � Car-boxylic acids Mix, 0.1 � Bu�er System 2, pH 7.5,50 % Precipitant Mix 1. Tery_3377-Cit data werecollected at ESRF beamline ID30B to resolution of1.2 Å.

The XDS package (60) was used to index,integrate and scale di�raction data. Structures havebeen determined by molecular replacement withPhaser (61) using coordinates of apo-FutA2 fromSynechocystis (PDB entry: 2VOZ, 21) and were fur-ther refined with REFMAC5 (62). During refine-ment, hydrogens were added in their riding positions,

and refinement was iterated with water-building andmodel correction in Coot (63).

Immunoblotting

Synechocystis cells were resuspended in 100 µl of10 m� Tris-HCl pH 7.3 and heated for 5 minutes at100 �C followed by centrifugation at 10,000 rpm for5 minutes at 4 �C. Samples were analyzed by SDS-PAGE on 4–12% NuPAGE acrylamide gradient gels(Life Technologies). Protein was blotted onto ni-trocellulose membrane, which was blocked for 1 hrin Tris-bu�ered saline with Tween 20 (TSB-T) and2% ECL Advance blocking reagent (GE Health-care). Antibody incubations (1 h) and membranewashes were performed in TBS-T. Anti-IdiA andanti-Slr0513 (FutA2) antisera were a kind gift fromE. Pistorius (64, 24), were used at concentrationsof 1:5000 and 1:10000 respectively. Following ap-plication of ECL Advance detection reagent (GEHealthcare) chemiluminescence was visualized us-ing a VersaDocTM CCD imager (Bio-Rad).

Acknowledgments: We thank Chris Holes of the macromolecular crystallization facility at the University ofSouthampton, sta� at the Diamond Light Source and the European Synchrotron Radiation Facility for accessand excellent user support through proposals MX11945 and MX1848, respectively. We thank sta� at theBiomedical Imaging Unit (BIU), Southampton General Hospital for assistance with confocal microscopy.

Conflict of interest: The authors declare that they have no conflicts of interest with the contents of thisarticle.

Author contributions: A.H. and T.S.B. conceived the study. D.P., M.M.M, A.H., T.S.B. and I.T. designedthe experiments. D.P., M.M.M., A.H., F.M. and I.T. performed the experiments. All authors interpretedresults and/or performed data analysis. D.P., M.M.M., A.H., T.S.B. and I.T. wrote the paper, which wasedited and approved by all authors.

9

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


REFERENCES1. Raven, J. A., Evans, M. C., and Korb, R. E. (1999) The role of trace metals in photosynthetic

electron transport in o2-evolving organisms, Photosynth. Res. 60(2-3), 111–150.2. Blaby-Haas, C. E. and Merchant, S. S. (2012) The ins and outs of algal metal transport, Biochimica

et Biophys. Acta (BBA)-Molecular Cell Res. 1823(9), 1531–1552.3. Shi, T., Sun, Y., and Falkowski, P. G. (2007) E�ects of iron limitation on the expression of metabolic

genes in the marine cyanobacterium trichodesmium erythraeum ims101, Environ. Microbiol. 9(12),2945–2956.

4. Richier, S., Macey, A. I., Pratt, N. J., Honey, D. J., Moore, C. M., and Bibby, T. S. (2012)Abundances of iron-binding photosynthetic and nitrogen-fixing proteins of trichodesmium both inculture and in situ from the north atlantic, PLOS One 7(5), e35571.

5. Snow, J. T., Polyviou, D., Skipp, P., Chrismas, N. A., Hitchcock, A., Geider, R., Moore, C. M.,and Bibby, T. S. (2015) Quantifying integrated proteomic responses to iron stress in the globallyimportant marine diazotroph trichodesmium, PLOS One 10(11), e0142626.

6. Moore, J. and Braucher, O. (2007) Observations of dissolved iron concentrations in the worldocean: implications and constraints for ocean biogeochemical models, Biogeosciences Discuss.4(2), 1241–1277.

7. Rubin, M., Berman-Frank, I., and Shaked, Y. (2011) Dust-and mineral-iron utilization by themarine dinitrogen-fixer trichodesmium, Nat. Geosci. 4(8), 529–534.

8. Hansel, C. M., Buchwald, C., Ossolinski, J., Dyhrman, S., Van Mooy, B., Polyviou, D., et al. (2016)Dynamics of extracellular superoxide production by trichodesmium colonies from the sargasso sea,Limnol. Oceanogr. 61(4), 1188–1200.

9. Falkowski, P. G. (1997) Evolution of the nitrogen cycle and its influence on the biological seques-tration of co2 in the ocean, Nat. 387(6630), 272.

10. Capone, D. G., Zehr, J. P., Paerl, H. W., Bergman, B., and Carpenter, E. J. (1997) Trichodesmium,a globally significant marine cyanobacterium, Sci. 276(5316), 1221–1229.

11. Westberry, T. K. and Siegel, D. A. (2006) Spatial and temporal distribution of trichodesmiumblooms in the world’s oceans, Glob. Biogeochem. Cycles 20(4).

12. Achilles, K. M., Church, T. M., Wilhelm, S. W., Luther, G. W., and Hutchins, D. A. (2003)Bioavailability of iron to trichodesmium colonies in the western subtropical atlantic ocean, Limnol.Oceanogr. 48(6), 2250–2255.

13. Webb, E. A., Mo�ett, J. W., and Waterbury, J. B. (2001) Iron stress in open-ocean cyanobacteria(synechococcus, trichodesmium, andcrocosphaera spp.): Identification of the idia protein, Appl.Environ. Microbiol. 67(12), 5444–5452.

14. Singh, A. K., McIntyre, L. M., and Sherman, L. A. (2003) Microarray analysis of the genome-wideresponse to iron deficiency and iron reconstitution in the cyanobacterium synechocystis sp. pcc6803, Plant Physiol. 132(4), 1825–1839.

15. Webb, E. A., Jakuba, R. W., Mo�ett, J. W., and Dyhrman, S. T. (2007) Molecular assessmentof phosphorus and iron physiology in trichodesmium populations from the western central andwestern south atlantic, Limnol. Oceanogr. 52(5), 2221–2232.

16. Rivers, A. R., Jakuba, R. W., and Webb, E. A. (2009) Iron stress genes in marine synechococcusand the development of a flow cytometric iron stress assay, Environ. Microbiol. 11(2), 382–396.

17. Chappell, P. D. and Webb, E. A. (2010) A molecular assessment of the iron stress response in thetwo phylogenetic clades of trichodesmium, Environ. Microbiol. 12(1), 13–27.

18. Michel, K.-P. and Pistorius, E. K. (2004) Adaptation of the photosynthetic electron transport chainin cyanobacteria to iron deficiency: the function of idia and isia, Physiol. Plantarum 120(1), 36–50.

19. Exss-Sonne, P., Tölle, J., Bader, K. P., Pistorius, E. K., and Michel, K.-P. (2000) The idia proteinof synechococcus sp. pcc 7942 functions in protecting the acceptor side of photosystem ii under

10

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


oxidative stress, Photosynth. Res. 63(2), 145–157.20. Katoh, H., Hagino, N., and Ogawa, T. (2001) Iron-binding activity of futa1 subunit of an abc-type

iron transporter in the cyanobacterium synechocystis sp. strain pcc 6803, Plant Cell Physiol. 42(8),823–827.

21. Badarau, A., Firbank, S. J., Waldron, K. J., Yanagisawa, S., Robinson, N. J., Banfield, M. J., andDennison, C. (2008) Futa2 is a ferric binding protein from synechocystis pcc 6803, J. Biol. Chem.283(18), 12520–12527.

22. Kranzler, C., Lis, H., Finkel, O. M., Schmetterer, G., Shaked, Y., and Keren, N. (2014) Coordinatedtransporter activity shapes high-a�nity iron acquisition in cyanobacteria, The ISME J. 8(2),409–417.

23. Fulda, S., Huang, F., Nilsson, F., Hagemann, M., and Norling, B. (2000) Proteomics of synechocys-tis sp. strain pcc 6803, Eur. J. Biochem. 267(19), 5900–5907.

24. Tölle, J., Michel, K.-P., Kruip, J., Kahmann, U., Preisfeld, A., and Pistorius, E. K. (2002) Local-ization and function of the idia homologue slr1295 in the cyanobacterium synechocystis sp. strainpcc 6803, Microbiol. 148(10), 3293–3305.

25. Waldron, K. J., Tottey, S., Yanagisawa, S., Dennison, C., and Robinson, N. J. (2007) A periplasmiciron-binding protein contributes toward inward copper supply, J. Biol. Chem. 282(6), 3837–3846.

26. Huang, F., Parmryd, I., Nilsson, F., Persson, A. L., Pakrasi, H. B., Andersson, B., and Norling, B.(2002) Proteomics of synechocystis sp. strain pcc 6803 identification of plasma membrane proteins,Mol. & Cell. Proteomics 1(12), 956–966.

27. Wang, S., Ogata, M., Horita, S., Ohtsuka, J., Nagata, K., and Tanokura, M. (2014) A novel modeof ferric ion coordination by the periplasmic ferric ion-binding subunit fbpa of an abc-type irontransporter from thermus thermophilus hb8, Acta Crystallogr. Sect. D: Biol. Crystallogr. 70(1),196–202.

28. Walworth, N., Pfreundt, U., Nelson, W. C., Mincer, T., Heidelberg, J. F., Fu, F., Waterbury, J. B.,del Rio, T. G., Goodwin, L., Kyrpides, N. C., et al. (2015) Trichodesmium genome maintainsabundant, widespread noncoding dna in situ, despite oligotrophic lifestyle, Proc. Natl. Acad. Sci.112(14), 4251–4256.

29. Spence, E., Sarcina, M., Ray, N., Møller, S. G., Mullineaux, C. W., and Robinson, C. (2003)Membrane-specific targeting of green fluorescent protein by the tat pathway in the cyanobacteriumsynechocystis pcc6803, Mol. Microbiol. 48(6), 1481–1489.

30. MacGregor-Chatwin, C., Sener, M., Barnett, S. F., Hitchcock, A., Barnhart-Dailey, M. C., Magh-laoui, K., Barber, J., Timlin, J. A., Schulten, K., and Hunter, C. N. (2017) Lateral segregation ofphotosystem i in cyanobacterial thylakoids, The Plant Cell 29(5), 1119–1136.

31. Kustka, A. B., Shaked, Y., Milligan, A. J., King, D. W., and Morel, F. M. (2005) Extracellularproduction of superoxide by marine diatoms: Contrasting e�ects on iron redox chemistry andbioavailability, Limnol. Oceanogr. 50(4), 1172–1180.

32. Shaked, Y., Kustka, A. B., and Morel, F. M. (2005) A general kinetic model for iron acquisition byeukaryotic phytoplankton, Limnol. Oceanogr. 50(3), 872–882.

33. Garg, S., Rose, A. L., Godrant, A., and Waite, T. D. (2007) Iron uptake by the ichthyotoxicchattonella marina (raphidophyceae): impact of superoxide generation1, J. Phycol. 43(5), 978–991.

34. Wang, S., Ogata, M., Horita, S., Ohtsuka, J., Nagata, K., and Tanokura, M. (2014) A novel modeof ferric ion coordination by the periplasmic ferric ion-binding subunit fbpa of an abc-type irontransporter from Thermus thermophilus hb8, Acta Crystallogr. Sect. D: Biol. Crystallogr. 70(1),196–202.

35. Berntsson, R. P.-A., Smits, S. H., Schmitt, L., Slotboom, D.-J., and Poolman, B. (2010) A structuralclassification of substrate-binding proteins, FEBS Lett. 584(12), 2606–2617.

11

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


36. Scheepers, G. H., Lycklama a Nijeholt, J. A., and Poolman, B. (2016) An updated structuralclassification of substrate-binding proteins, FEBS Lett. 590(23), 4393–4401.

37. Tom-Yew, S. A., Cui, D. T., Bekker, E. G., and Murphy, M. E. (2005) Anion-independent ironcoordination by the Campylobacter jejuni ferric binding protein, J. Biol. Chem. 280(10), 9283–9290.

38. Koropatkin, N., Randich, A. M., Bhattacharyya-Pakrasi, M., Pakrasi, H. B., and Smith, T. J. (2007)The structure of the iron-binding protein, futa1, from Synechocystis 6803, J. Biol. Chem. 282(37),27468–27477.

39. Shouldice, S. R., Skene, R. J., Dougan, D. R., Snell, G., McRee, D. E., Schryvers, A. B., andTari, L. W. (2004) Structural basis for iron binding and release by a novel class of periplasmiciron-binding proteins found in gram-negative pathogens, J. Bacteriol. 186(12), 3903–3910.

40. Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990) Basic local alignmentsearch tool, J. Mol. Biol. 215(3), 403–410.

41. Li, W. and Godzik, A. (2006) Cd-hit: a fast program for clustering and comparing large sets ofprotein or nucleotide sequences, Bioinforma. 22(13), 1658–1659.

42. Dereeper, A., Guignon, V., Blanc, G., Audic, S., Bu�et, S., Chevenet, F., Dufayard, J.-F., Guindon,S., Lefort, V., Lescot, M., et al. (2008) Phylogeny. fr: robust phylogenetic analysis for the non-specialist, Nucleic acids research 36(suppl 2), W465–W469.

43. Dereeper, A., Audic, S., Claverie, J.-M., and Blanc, G. (2010) Blast-explorer helps you buildingdatasets for phylogenetic analysis, BMC Evol. Biol. 10(1), 8.

44. Notredame, C., Higgins, D. G., and Heringa, J. (2000) T-co�ee: A novel method for fast andaccurate multiple sequence alignment, J. Mol. Biol. 302(1), 205–217.

45. Castresana, J. (2000) Selection of conserved blocks from multiple alignments for their use inphylogenetic analysis, Mol. Biol. Evol. 17(4), 540–552.

46. Guindon, S. and Gascuel, O. (2003) A simple, fast, and accurate algorithm to estimate largephylogenies by maximum likelihood, Syst. Biol. 52(5), 696–704.

47. Anisimova, M. and Gascuel, O. (2006) Approximate likelihood-ratio test for branches: A fast,accurate, and powerful alternative, Syst. Biol. 55(4), 539–552.

48. Rippka, R., Deruelles, J., Waterbury, J. B., Herdman, M., and Stanier, R. Y. (1979) Genericassignments, strain histories and properties of pure cultures of cyanobacteria, Microbiol. 111(1),1–61.

49. Shcolnick, S., Shaked, Y., and Keren, N. (2007) A role for mrga, a dps family protein, in theinternal transport of fe in the cyanobacterium synechocystis sp. pcc6803, Biochimica et Biophys.Acta (BBA)-Bioenergetics 1767(6), 814–819.

50. Kranzler, C., Lis, H., Shaked, Y., and Keren, N. (2011) The role of reduction in iron uptakeprocesses in a unicellular, planktonic cyanobacterium, Environ. Microbiol. 13(11), 2990–2999.

51. Kolber, Z. S., Práöil, O., and Falkowski, P. G. (1998) Measurements of variable chlorophyll fluo-rescence using fast repetition rate techniques: defining methodology and experimental protocols,Biochimica et Biophys. Acta (BBA)-Bioenergetics 1367(1), 88–106.

52. Thomas, J. D., Daniel, R. A., Errington, J., and Robinson, C. (2001) Export of active greenfluorescent protein to the periplasm by the twin-arginine translocase (tat) pathway in escherichiacoli, Mol. Microbiol. 39(1), 47–53.

53. Bendtsen, J. D., Nielsen, H., Widdick, D., Palmer, T., and Brunak, S. (2005) Prediction of twin-arginine signal peptides, BMC bioinformatics 6(1), 167.

54. Petersen, T. N., Brunak, S., von Heijne, G., and Nielsen, H. (2011) Signalp 4.0: discriminatingsignal peptides from transmembrane regions, Nat. Methods 8(10), 785–786.

55. Pédelacq, J.-D., Cabantous, S., Tran, T., Terwilliger, T. C., and Waldo, G. S. (2006) Engineeringand characterization of a superfolder green fluorescent protein, Nat. Biotechnol. 24(1), 79–88.

12

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


56. Hollingshead, S., Kope�ná, J., Jackson, P. J., Canni�e, D. P., Davison, P. A., Dickman, M. J.,Sobotka, R., and Hunter, C. N. (2012) Conserved chloroplast open-reading frame ycf54 is requiredfor activity of the magnesium protoporphyrin monomethylester oxidative cyclase in synechocystispcc 6803, J. Biol. Chem. 287(33), 27823–27833.

57. Williams, J. G. (1988) [85] construction of specific mutations in photosystem ii photosyntheticreaction center by genetic engineering methods in synechocystis 6803, Methods Enzymol. 167,766–778.

58. Cereda, A., Hitchcock, A., Symes, M. D., Cronin, L., Bibby, T. S., and Jones, A. K. (2014) Abioelectrochemical approach to characterize extracellular electron transfer by synechocystis sp.pcc6803, PLOS One 9(3), e91484.

59. Aldridge, C., Spence, E., Kirkilionis, M. A., Frigerio, L., and Robinson, C. (2008) Tat-dependenttargeting of rieske iron-sulphur proteins to both the plasma and thylakoid membranes in thecyanobacterium synechocystis pcc6803, Mol. Microbiol. 70(1), 140–150.

60. Kabsch, W. (2010) Xds, Acta Crystallogr. Sect. D: Biol. Crystallogr. 66(2), 125–132.61. McCoy, A. J., Grosse-Kunstleve, R. W., Adams, P. D., Winn, M. D., Storoni, L. C., and Read, R. J.

(2007) Phaser crystallographic software, J. Appl. Crystallogr. 40(4), 658–674.62. Murshudov, G. N., Vagin, A. A., and Dodson, E. J. (1997) Refinement of macromolecular structures

by the maximum-likelihood method, Acta Crystallogr. Sect. D: Biol. Crystallogr. 53(3), 240–255.63. Emsley, P. and Cowtan, K. (2004) Coot: model-building tools for molecular graphics, Acta

Crystallogr. Sect. D: Biol. Crystallogr. 60(12), 2126–2132.64. Michel, K.-P. and Pistorius, E. K. (1992) Isolation of a photosystem ii associated 36 kda polypep-

tide and an iron-stress 34 kda polypeptide from thylakoid membranes of the cyanobacteriumsynechococcus pcc 6301 grown under mild iron deficiency, Zeitschrift für Naturforschung C47(11-12), 867–874.

13

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


FOOTNOTESDespo Polyviou was funded by a studentship from NERC, the Graduate School of the National OceanographyCentre Southampton (GSNOCS) and A.G. Leventis Foundation. Moritz M. Machelett was funded by astudentship from the National Oceanography Centre and Biological Sciences, University of Southampton.

Abbreviations: ABC, ATP-binding cassette; EPR, electron paramagnetic resonance; FBP, ferric bindingprotein; FZ, ferrozine; IdiA, iron deficiency induced protein A; PSII, photosystem II.

TABLES

Table 1: Crystallographic analysis: data collection and refinement statisticsValues in parentheses refer to the highest resolution shell.

Tery_3377-Ala Tery_3377-Water Tery_3377-CitPDB accession code 6G7N 6G7P 6G7Q

Data collectionSpace group P21 P21 P21Cell dimensions

a, b, c [Å] 56.2, 83.2, 65.2 56.0, 82.9, 64.9 56.3, 102.0, 57.7↵, �, � [�] 90, 95.20, 90 90, 95.20, 90 90, 112.85, 90

Resolution [Å] 29.26-1.1 (1.130-1.1) 29.11-1.5 (1.540-1.5) 28.64-1.2 (1.230-1.2)Rmerge [%] 0.04759 (0.3528) 0.04115 (0.2744) 0.07207 (0.8711)I/� (I) 11.83 (3.00) 16.22 (4.13) 11.63 (1.55)CC 1/2 0.998 (0.765) 0.999 (0.917) 0.999 (0.679)Completeness [%] 99.42 (99.56) 95.48 (94.11) 98.30 (99.17)Redundancy 3.2 (3.0) 3.5 (3.6) 5.8 (5.6)

RefinementResolution [Å] 29.26-1.1 (1.129-1.1) 29.11-1.5 (1.539-1.5) 28.58-1.2 (1.231-1.2)No. reflections 227890 (16772) 85685 (6274) 181790 (13467)Rwork [%] 0.1337 (0.2358) 0.1476 (0.1790) 0.1355 (0.2630)R f ree [%] 0.1512 (0.2377) 0.1739 (0.2050) 0.1648 (0.2840)Monomers/AU 2 2 2No. of atoms 6236 5656 6198

Protein 5250 5106 5193Ligand/ion 24 11 49Water 962 539 956

B-factors 16.42 18.12 17.88Protein 15.24 17.90 16.46Ligand/ion 27.24 34.35 26.75Water 29.35 24.75 31.17

R.m.s. deviationsBond length [Å] 0.012 0.018 0.019Bond angle [�] 1.57 1.89 2.01

14

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


FIGURES

Figure 1: Maximum likelihood phylogenetic analysis of Tery_3377 homologs, based on amino acidsequence. homologs of Tery_3377 and Synechocystis FutA1 and FutA2 were identified using BlastP (40)and aligned along with sequences from the cyanobacteria Synechococcus elongatus PCC 6301, Prochloro-coccus marinus MIT 9301, and Anabaena PCC 7120, and the pathogenic bacteria Haemophilus influenzaand Neisseria meningitides. The graphical representation was produced using FigTree v1.4.3 and branchprobabilities reported are calculated using an approximate likelihood ratio test (aLRT, 47). The SynechocystisFutA1 (red) and FutA2 (blue) proteins are separated into two groups, and Tery_3377 clusters within theFutA1-like group. Filled circles represent branches of the tree for which there is experimental evidence ofintracellular (red) or Fe transport (blue) function. Database accessions for each protein (SwissProt wherepossible, otherwise RefSeq) are given in brackets. The scale bar represents the number of substitutions persite.

15

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Figure 2: Localization of signal sequence-GFP constructs in Synechocystis. Using confocal microscopythe WT (A) was compared to strains expressing GFP fused to the signal sequences of the E. coli controlTorA (B), Synechocystis FutA1 (C), Synechocystis FutA2 (D) and Trichodesmium Tery_3377 (E); the GFPcontrol, i.e. without signal sequence, is shown in (F). Imaging was performed under identical settings withexcitation at 488 nm and is shown as GFP emission at 510-530 nm (i) and as autofluorescence (representedby phycocyanin) at 600-700 nm (ii); the merged images are shown in (iii). Periplasmic localization of GFPis indicated by green fluorescent halos around the periphery of the cells in all strains (B-E) but not in WT(A) or the GFP control (F), the latter showing cytosolic expression in absence of a signal peptide.

16

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Figure 3: Reduction in growth and photosynthetic e�ciency of strains grown in the presence offerrozine. The decrease in growth rates measured as µ (day-1) (A) and photosynthetic e�ciency measuredas Fv/Fm (B) during exponential growth of Synechocystis in YBG11 media when grown with and withoutthe Fe2+ binding ligand ferrozine (FZ). Data are presented for WT, �futA1 and �futA2 mutants and the samestrains expressing the Trichodesmium tery_3377 gene (�futA1+ and �futA2+). Error bars show the standarddeviation from the mean of three biological replicates. Statistical significance (*) was assessed using theStudent’s t-test (p < 0.05).

17

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Figure 4: Electron paramagnetic resonance spectrum (EPR) of reconstituted Tery_3377 probing theiron centre. Recombinant purified Tery_3377 from Trichodesmium erythraeum ISM101 produced in E. coliwas incubated with Ammonium Fe(II) sulfate under aerobic conditions; the complex was repurified by sizeexclusion chromatography to remove unbound Fe. The EPR spectrum was collected from a sample with0.55 m� protein concentration.

18

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Figure 5: Structure analysis of Tery_3377. (A) Iron-binding site of Tery_3377-Ala with amino acid(ala) ligand (left); the electron-density map shown (middle) is obtained after molecular replacement andsubsequent refinement without ligands (FO - FC, contoured at 3 sigma); ligand coordination (right) shows ironin a trigonal bi-pyramidal geometry; red: oxygen, blue: nitrogen, brown sphere: ferric-ion, yellow dashes:h-bonds to axial coordination sites. (B) Iron-binding site of Tery_3377-Water with water (W) ligands (left);the electron-density map shown (middle) is obtained after molecular replacement and subsequent refinementwithout ligands (Fo - Fc, contoured at 3 sigma); ligand coordination (right) shows iron in an octahedralgeometry; red: oxygen, brown sphere: ferric-ion, gray sphere: water, yellow dashes: h-bonds to axialcoordination sites.

19

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Figure 6: Iron binding site and coordination of citrate molecules. (A) The structure of Tery_3377-Cit.The fold is illustrated by the yellow ribbon, and the iron-binding site is shown by sticks representing thetyrosines shown in figure 5 (yellow) and two citrate molecules (blue). The iron coordinated by the two citratemolecules is shown as a brown sphere. The magenta circle indicates the position an iron ligand would occupyin iron-loaded Tery_3377. (B) Several protein amino acids, shown as yellow sticks, coordinate the citrateresidues, shown in blue. The electron-density map is obtained after molecular replacement and subsequentrefinement without ligands (FO - FC, contoured at 2 sigma). Red: oxygen, brown: ferric-ion, gray-mesh:electron density, blue dashes: coordinating interactions.

20

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Figure 7: Analysis of iron coordination in Tery_3377. (A) The open state of FutA2 from Synechocystis sp.PCC6803 is Fe free (gray). When superposed with the closed Fe bound form of FutA2 from Synechocystissp. PCC6803 (Fe-FutA2), the C-terminal domains can be brought into perfect overlap while the N-terminaldomains reveal structural changes (green and light green) indicative of a closing of the substrate bindingcleft. Notably, a loop region contributing binding residues to the Fe binding sites is observed in a di�erentposition (green) to the open state. (B) Superposition of Tery_3377-Ala (blue) with Fe-FutA2 (green). Theloop from the N-terminal domain brings Fe co-ordinating tyrosine and histidine side chains into contactdistance, allowing Fe coordination. (C) Coordination site of superposed Tery_3377-Ala (blue) with Fe-FutA2 (green). The equivalent amino acid residues to His13 and Tyr14, His43 and Tyr44 in Tery_3377, arenot seen in contact with iron in either Tery_3377-Ala nor Tery_3377-Water. However, the polar atoms fromthe alanine ligand in Tery_3377-Ala superpose well with the coordinating atoms of His13 and Tyr14. Thethree co-ordinating tyrosine residues originating from the C-terminal domains occupy the same position inTery_3377-Ala and Fe-FutA2. Coordinating residues (shown as sticks) and ferric iron (brown sphere).

21

at UN

IV O

F EA

ST A

NG

LIA

on October 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from
http://www.jbc.org/

structural and functional characterisation of idia/futa (tery ......structure and function of...

Documents